Planet enlists machine learning experts to parse Amazon basin data | Hacker News Reader

Planet enlists machine learning experts to parse Amazon basin data

techcrunch.com

104 pointsrobinkraft21 comments

Unfortunately, the Amazon won't be saved unless the rest of the world puts their hand in their pockets. It is a resource for those countries to elevate themselves out of poverty and develop their countries.

Europe did the same. The US did the same. To tell these poor countries they should not cut down the forest because it is a resource for the world is a bit rich.

Yes there is a problem with the mega farms / ranches that seem to be a problem in some areas of Brazil, but also the small holdings where peasant (I don't like to use this word, but it does differentiate from the typical wesstern usage of the word farmer, but I don't intend to use it in a derogatory way) farmers eek out an existence by chopping down a few hectares for pasture land for a couple dozen head of cattle which they can sell.

Ecuador tried a novel approach a couple of years ago and went to the UN asking for $4 billion to not extract oil from a particular life rich area of the forest (the oil was said to be worth $8billion). They got pledges of $32 million (pledges which do not always translate to payment).

So now oil roads are being built, where the road goes, small holidings sprint up all along and more of the forest gets cut down / polluted.

I live so much in the bubble that whenever I read "Amazon Rainforest" I think about what new cloud product by AWS that is...

Same. I thought article is about how AWS machine learning platform found a big customer which would help the revenue growth

https://www.kaggle.com/c/planet-understanding-the-amazon-fro...

Thanks @ontouchstart

Do you get extra rewards if you score well with random forests ?

Or if you use AWS?

Does anyone else believe that chopping down massive swaths of rainforest is the primary cause of global warming?

I mean, these are the planets largest carbon sinks and they are being decimated.

Definitely not the primary cause of global warming.

Speaking terms of carbon flows, the Amazon as a whole only absorbs two or three times the carbon emissions as the UK.

However it's approaching saturation anyway--it can't absorb an infinite amount of carbon.

In terms of absolute numbers, the total carbon locked up in the Amazon is definitely large, but I'd guess it's at least an order of magnitude less than human carbon emissions since the start of industrialization. Trying to figure out how to get the exact figure.

Some back of napkin numbers:

Rainforests store 250 tons carbon/hectare

The Amazon is 550,000,000 hectares in area

The Amazon has 137 gigatons of C total

Since 1850, humans have released around 1500 gigatons.

In particular, recently from 2000 to 2015, around 500 gigatons have been released by human activity.

Around 2700 gigatons remain in the ground.

All of these statements are just random numbers I found off the Internet, so take them with a gigaton of sodium.

Not primary, but deforestation is still significant. From Scientific American:

"By most accounts, deforestation in tropical rainforests adds more carbon dioxide to the atmosphere than the sum total of cars and trucks on the world’s roads. According to the World Carfree Network (WCN), cars and trucks account for about 14 percent of global carbon emissions, while most analysts attribute upwards of 15 percent to deforestation."

https://www.scientificamerican.com/article/deforestation-and...

> the Amazon as a whole only absorbs two or three times the carbon emissions as the UK.

The Amazon probably doesn't absorb any net amount of carbon, unless it's growing.

Let's demystify this: The carbon trapped in the Amazon is the plants. When plants consume CO₂, the C part becomes the plant. In a forest in equilibrium, growing plants do capture carbon from the air, but rotting dead plants release an equal amount.

So when you chop down a forest, and for the sake of example burn all the plants, you're releasing a large one-time amount of CO₂, but then the whole damage is done. And if you then leave the area alone, the forest will regenerate over the next 50-100 years, reabsorb the same amount of CO₂.

If you chop down the forest and build lasting structures from the wood, you're actually removing net CO₂!

I wonder if we could come up with some sort of fast growing plant that we could then use as a replacement for sand as a filler. This would reduce the sand crisis and capture carbon, if it's possible.

There is a sand crisis too?

Is it the one caused by dams?

Between 2011 and 2015 China poured more concrete than the US did in the entire 20th century.

Sand is one component of most concrete.

https://www.theguardian.com/cities/2017/feb/27/sand-mining-g...

See https://news.ycombinator.com/item?id=13789108

The old American indigenous civilizations were way more prolific with the slashing and the burning than we are.

"It's the industry, stupid" - William Jefferson Clinton 2017

From what I heard, peat is a much bigger carbon sink. I mean, if you have multiple meter thick layers of practically pure carbon material covering the land, thats an enourmous amount. I have read something about 30% of carbon stored on 3% of the worlds land mass. The amazon on the contrary is only a very thin layer of organic soil with trees on top. Its still huge but only through its size, not by amout per area.

$60,000, that's it?

Resulting models will likely be released as open source, and the techniques published in open journals.

The big money is in productionizing and running it on bigger datasets.

Kaggle competitions are not a data-science-product-for-hire kind of thing, like some kind of logo design contest. It is a sport. Super GM chess competitions see smaller prize pools.

Mediocre people, including me, would compete for free (just to get access to interesting data like this). The really talented people are not driven by money, but by competition and fame.

Increasing the prize money is more of a marketing move and only attracts more low-to-mediocre people trying to get lucky in a lottery: It won't increase the quality of top 10 solutions. And no computer vision PhD/professor is going to drop everything he/she is working on for a small chance to win 400k.

If it's as you say, then OK.

But it did look like a fishing expedition for ideas worth far in excess of $60,000, which smelled exploitative

Reply on news.ycombinator.com